<p>This study investigates the influence of ball-end micromill geometry, feed per tooth, and workpiece microstructure on the surface integrity and dimensional accuracy in the micro-milling of low-carbon steels. Experiments were performed on dual-phase (DPh) and ultrafine-grained (UFG) steels, complemented by statistical analyses and machine learning models. The results showed that tool deflection affected profile fidelity, with the diameter-neck length ratio being more decisive than the tool size alone. Lower feeds promoted smoother surfaces, whereas higher feeds increased roughness and geometric deviations, particularly in up-milling. Microstructural effects were also observed; UFG steel enabled more stable and predictable roughness formation, whereas DPh steel exhibited better dimensional consistency. The burrs were minimal across all conditions owing to the favorable edge angles inherent to the tool-workpiece interaction. Predictive modeling confirmed greater accuracy for UFG steel (<i>R</i>² ≈ 0.71), highlighting the role of homogeneous microstructures in enhancing the roughness prediction. Overall, the study provides new insights into the interplay between tool geometry, feed strategy, and material microstructure to support process planning in micro-milling of low-cabon steels.</p>